[ltp] My war with T40 and ACPI drain

[James McKenzie]

Peter Frühberger wrote:
> Bob Alexander wrote:
> 
>> Peter Frühberger wrote:
>>
>>> Bob Alexander wrote:
>>>
>>>> Peter Frühberger wrote:
>>>>
>>>>> Hi Bob, try this link,
>>>>> http://www.stud.uni-karlsruhe.de/~unatk/media/radeonstuff.tar.bz2
>>>>> here I got it working with 2.6.11-rc2,
>>>>> just read the info file, change the Power File to add your model 
>>>>> and apply the patches in the given order.
>>>>> If you want you can also apply swsusp2-2.1.5.15 for 2.6.11-rc2
>>>>>
>>>>> Good Luck
>>>>> Peter
>>>>>
>>>>>
>>>>
>>>> Peter,
>>>> (and of course all other friends giving me some hope and support).
>>>>
>>>> Let me understand better before I act.
>>>>
>>>> I currently have a 2.6.11-rc2 with ibm_acpi-0.10, swsusp2 
>>>> 2.1.5.15(and a patch to this).
>>>>
>>>> If I understand correctly I shoud edit the 
>>>> 05-radeonfb-Thinkpad-Power.patch file by REPLACING/OVERWRITING the 
>>>> first .ident entry with proper values for my thinkpad (which I can 
>>>> see also with the dmidecode command) and the corresponding .matches 
>>>> entry.
>>>>
>>>> I will then apply the patches in order ... did I get it ? :)
>>>>
>>>> TIA,
>>>> Bob
>>>
>>>
>>>
>>>
>>> Yes this is the way to go, for example the entry for my R40 2722-B3G 
>>> looks like that:
>>>
>>> +       {
>>> +               .ident = "IBM ThinkPad R40 (2722-B3G)",
>>> +               .matches = {
>>> +                       DMI_MATCH(DMI_SYS_VENDOR, "IBM"),
>>> +                       DMI_MATCH(DMI_PRODUCT_NAME, "2722B3G"),
>>> +               },
>>> +       },
>>>
>>> just edit this entry to fit to your values.
>>>
>>> If you go into S3 mode you should see sth. like that:
>>> radeonfb (0000:01:00.0): suspending to state: 3...
>>> radeonfb: switching to D2 state...
>>> radeonfb (0000:01:00.0): resuming from state: 3...
>>> radeonfb: switching to D0 state...
>>>
>>> D2 state is the important thing.
>>>
>>> I did compile radeonfb (new version) directly into the kernel and my 
>>> append lilo conf for 2.6.11-rc2 looks like that:
>>>         append="video=radeonfb:1024x768-16@60 apm=off acpi=on 
>>> pci=routeirq resume2=swap:/dev/hda5 acpi_sleep=s3_bios"
>>>
>>> good luck
>>> Peter
>>>
>>>
>>>
>> Peter,
>> thank you for the ongoing support. Still no luck.
>>
>> I added radeonfb in my /etc/modules since I could not see any messages 
>> related to radeonfb in my logs.
>>
>> I also added the same kern params in my GRUB menu.lst config (only 
>> diff is video=radeonfb:1400x1050-24@60 for my screen).
>>
>> Now that I have done:
>>
>> a) During boot the screen stays black for quite a long time (I finally 
>> see the last few lines of the console boot before GDM kicks in)
>>
>> b) When I suspend I do not see the D2 state (see below)
>>
>> c) The power drain is still in the 4.8W per hour :( :( :(
>>
>>
>> Here is some data:
>>
> 
> I`ll send you a deb image which you could try to test...
> offline
> 
> 
> 
This may come as a surprise to the list, but I have an A22p and decided 
to get a patch so I could suspend/hibernate using ACPI.  This system is 
affected by the 'power drain' problem.  If I suspend under APM, the 
system will be able to sit for about 14 days.  If I suspend under ACPI, 
and after about four hours, I lost 75% of battery capacity.  I thought 
that the problem was isolated to systems with ATI video cards, but it 
looks like it might be all ATI cards.  One thing that I noticed is that 
this problem might be associated with the DRI module.  Is there a patch 
for DRI?

James McKenzie